RESEARCH ARTICLE Open Access

Ten new complete mitochondrial genomes ofpulmonates (Mollusca: Gastropoda) and theirimpact on phylogenetic relationshipsTracy R White1, Michele M Conrad1, Roger Tseng1, Shaina Balayan1, Rosemary Golding2,António Manuel de Frias Martins3 and Benoît A Dayrat1*

Abstract

Background: Reconstructing the higher relationships of pulmonate gastropods has been difficult. The use ofmorphology is problematic due to high homoplasy. Molecular studies have suffered from low taxon sampling.Forty-eight complete mitochondrial genomes are available for gastropods, ten of which are pulmonates. Here arepresented the new complete mitochondrial genomes of the ten following species of pulmonates: Salinatorrhamphidia (Amphiboloidea); Auriculinella bidentata, Myosotella myosotis, Ovatella vulcani, and Pedipes pedipes(Ellobiidae); Peronia peronii (Onchidiidae); Siphonaria gigas (Siphonariidae); Succinea putris (Stylommatophora);Trimusculus reticulatus (Trimusculidae); and Rhopalocaulis grandidieri (Veronicellidae). Also, 94 new pulmonate-specific primers across the entire mitochondrial genome are provided, which were designed for amplifying entiremitochondrial genomes through short reactions and closing gaps after shotgun sequencing.

Results: The structural features of the 10 new mitochondrial genomes are provided. All genomes share similargene orders. Phylogenetic analyses were performed including the 10 new genomes and 17 genomes fromGenbank (outgroups, opisthobranchs, and other pulmonates). Bayesian Inference and Maximum Likelihoodanalyses, based on the concatenated amino-acid sequences of the 13 protein-coding genes, produced the sametopology. The pulmonates are paraphyletic and basal to the opisthobranchs that are monophyletic at the tip ofthe tree. Siphonaria, traditionally regarded as a basal pulmonate, is nested within opisthobranchs. Pyramidella,traditionally regarded as a basal (non-euthyneuran) heterobranch, is nested within pulmonates. Several hypothesesare rejected, such as the Systellommatophora, Geophila, and Eupulmonata. The Ellobiidae is polyphyletic, but thefalse limpet Trimusculus reticulatus is closely related to some ellobiids.

Conclusions: Despite recent efforts for increasing the taxon sampling in euthyneuran (opisthobranchs andpulmonates) molecular phylogenies, several of the deeper nodes are still uncertain, because of low support values aswell as some incongruence between analyses based on complete mitochondrial genomes and those based onindividual genes (18S, 28S, 16S, CO1). Additional complete genomes are needed for pulmonates (especially forWilliamia, Otina, and Smeagol), as well as basal heterobranchs closely related to euthyneurans. Increasing the numberof markers for gastropod (and more broadly mollusk) phylogenetics also is necessary in order to resolve some of thedeeper nodes -although clearly not an easy task. Step by step, however, new relationships are being unveiled, suchas the close relationships between the false limpet Trimusculus and ellobiids, the nesting of pyramidelloids withinpulmonates, and the close relationships of Siphonaria to sacoglossan opisthobranchs. The additional genomespresented here show that some species share an identical mitochondrial gene order due to convergence.

* Correspondence: bdayrat@ucmerced.edu1School of Natural Sciences, University of California, 5200 North Lake Road,Merced, CA 95343, USAFull list of author information is available at the end of the article

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

© 2011 White et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.

BackgroundElucidating the higher phylogenetic relationships of pul-monate gastropods has remained difficult. A morphol-ogy-based phylogenetic analysis revealed a high level ofhomoplasy and resulted in a poorly-resolved tree [1].Molecular studies have been based on few individualgenes, essentially 18S, 28S, 16S, and COI data [2-5], orfew complete mitochondrial genomes [6-8].Analyses based on individual gene sequences all provide

similar relationships (Figure 1) [2-5]: pulmonates aremonophyletic, but they include a few taxa not traditionallyclassified as pulmonates (Acochlidia, traditionally regardedas opisthobranchs, and Glacidorbidae and Pyramidelloidea,traditionally regarded as basal heterobranchs); also,opisthobranchs are paraphyletic, basal to pulmonates;finally, the false limpet Siphonaria, traditionally regardedas a pulmonate, is in some cases found to be more closelyrelated to opisthobranchs than pulmonates. In recentyears, taxon sampling has significantly increased in ana-lyses utilizing individual genes: the largest data set (18S,16S, and COI) established so far includes 79 species repre-senting all major taxa of pulmonates [5].Analyses based on complete mitochondrial genomes

provide different phylogenetic relationships, at least forthe deep nodes [6-8]: pulmonates are paraphyletic, basalto the monophyletic opisthobranchs; Siphonaria is nestedwithin the opisthobranchs. Taxon sampling is still limitedin analyses based on complete mitochondrial genomes,mainly because gastropod mitochondrial genomes arestill difficult to obtain. Since the first complete gastropod

mitochondrial genome was published in 1995 [9], 48complete genomes have been made available (Figure 2).The use of shotgun sequencing and the decrease insequencing costs caused a noticeable increase in the pro-duction of gastropod mitochondrial genomes a few yearsago (Figure 2). Ten genomes became available in ninedifferent publications between 1995 and 2006; since2008, 38 genomes became available, 30 of whichappeared in only four publications [6,8,10,11], although afew papers with only one genome were also published.At present, Opisthobranchia and Neogastropoda (one

of the lineages of Caenogastropoda) are the taxa forwhich the largest number of complete mitochondrialgenomes are available, with 17 and 12 genomes, respec-tively (Figure 2). Ten genomes are available for pulmo-nates (including two unpublished), but only some of thepulmonate higher clades are represented. The taxonsampling for the other gastropods is either insufficient(Patellogastropoda, Vetigastropoda, Neritimorpha, basalcaenogastropods, and basal heterobranchs) or missing(Cocculiniformia).In the present study, we report 10 new, complete,

mitochondrial genomes of pulmonates, with a specialfocus on lineages that were poorly or not sampled(Figure 2, Table 1): Salinator rhamphidia (Amphiboloi-dea); Auriculinella bidentata, Myosotella myosotis, Ova-tella vulcani, and Pedipes pedipes (Ellobiidae); Peroniaperonii (Onchidiidae); Siphonaria gigas (Siphonariidae);Succinea putris (Stylommatophora); Trimusculus reticu-latus (Trimusculidae); and Rhopalocaulis grandidieri

Figure 1 Phylogenetic relationships of pulmonates. Relationships obtained in a recent study based on individual genes (18S, 16S, COI),including 79 pulmonate species [5]; the asterisk (*) indicates that the clade Ellobiidae includes three taxa that have not been traditionallyregarded as ellobiids (Otina, Smeagol, and Trimusculus). Node support values are cited using the following format: ‘’1.00/77’’ means that BIposterior probability = 1.00, and that ML bootstrap value = 77%. Only BI posterior probabilities > 0.75 and ML bootstrap values > 50% areshown.

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 2 of 15

(Veronicellidae). Here we also provide a set of 94 new,pulmonate-specific primers spanning the entire mito-chondrial genome and specifically designed for thepresent study. These new primers were combined

in multiple pairs, or with 10 mitochondrial primerspreviously published, to amplify genomes throughsimultaneous, short reactions. This direct approach wasof great help for amplifying mitochondrial genomes in

Figure 2 List of complete gastropod mitochondrial genomes available at present. The hypothesis of phylogenetic relationships is basedon both morphological and molecular data [46]. The reference in which each genome was made available [6-11,47-58] is indicated in bracketsas well as the year of publication.

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 3 of 15

the present study. Pulmonate relationships are evaluatedthrough phylogenetic analyses based on these new gen-omes as well as 17 genomes published previously. Theimpact of this new data set on our understanding of therelationships and evolution of pulmonates is discussed.

ResultsGenome structural featuresThe structural features of each of the 10 mitochondrialgenomes sequenced in this study are summarized inTable 2. Each genome consists of 13 protein-codinggenes, two rRNA genes, and 22 tRNA genes. The gen-omes vary in size from 13, 968 bp (Peronia peronii) to 16,708 bp (Pedipes pedipes) with most in the range of 14,000 bp. In all 10 genomes, 13 of the 37 genes are codedon the minus strand: trnQ, trnL2, atp8, trnN, atp6, trnR,trnE, rrnS, trnM, nad3, trnS2, trnT, and cox3. In Succineaputris, trnY and trnW are also coded on the minusstrand, as well as trnH in Siphonaria gigas. In the major-ity of the protein-coding genes (59 of 130), the start

codon is TTG. Alternatively, the start codon is eitherATG (43), GTG (17), ATT (6), ATA (3), CTG (1), orTTA (1). The stop codon is either TAA (52), TAG (41),T (36), or TA (1), none of which is used solely for a parti-cular protein-coding gene.All 10 mitochondrial genomes have adjacent overlap-

ping genes, typically 11 to 12 genes; the highest number(15) of overlapping genes is found in Peronia peronii andthe fewest number (7) is found in Pedipes pedipes, whichis likely correlated with genome size. The amount ofoverlap between genes is typically between 1 and 30 bp.Only one pair of overlapping genes is common to all 10mitochondrial genomes: trnK, cox1 (5 to 8 bp). With theexception of Pedipes pedipes, the other nine mitochon-drial genomes all have nad6, nad5 (2 to 18 bp) overlap-ping and nad5, nad1 (14 to 26 bp) overlapping. Thelargest overlaps in adjacent genes are 39 bp betweennad2, trnK in Succinea putris and 45 bp between nad4L,cob in Auriculinella bidentata. The number of intergenicspacers range from four in Peronia peronii to 20 in

Table 1 List of the species included in the present study

Classification Species Name Locality Voucher Genbank

Caenogastropoda Cymatium parthenopeum - - EU827200

Caenogastropoda Ilyanassa obsoleta - - DQ238598

Caenogastropoda Lophiotoma cerithiformis - - DQ284754

Basal Heterobranchia Pyramidella dolabrata - - AY345054

Pulmonata, Amphibolidae Salinator rhamphidia Australia, NSW CASIZ 180470 JN620539*

Pulmonata, Ellobiidae Auriculinella bidentata Azores CASIZ 184730 JN606066*

Pulmonata, Ellobiidae Myosotella myosotis Azores CASIZ 184731 JN606067*

Pulmonata, Ellobiidae Myosotella myosotis - - AY345053

Pulmonata, Ellobiidae Ovatella vulcani Azores CASIZ 180486 JN615139*

Pulmonata, Ellobiidae Pedipes pedipes Azores CASIZ 180476 JN615140*

Pulmonata, Hygrophila Radix balthica - - HQ330989

Pulmonata, Hygrophila Biomphalaria glabrata - - AY380531

Pulmonata, Onchidiidae Onchidella borealis - - DQ991936

Pulmonata, Onchidiidae Onchidella celtica - - AY345048

Pulmonata, Onchidiidae Peronia peronii Guam CASIZ 180486 JN619346*

Pulmonata, Siphonariidae Siphonaria gigas Panama UF 359645 JN627205*

Pulmonata, Siphonariidae Siphonaria pectinata - - AY345049

Pulmonata, Stylommatophora Albinaria coerulea - - X83390

Pulmonata, Stylommatophora Succinea putris France CASIZ 180491 JN627206*

Pulmonata Trimusculidae Trimusculus reticulatus California CASIZ 177988 JN632509*

Pulmonata, Veronicellidae Rhopalocaulis grandidieri Madagascar NM L7086 JN6193467*

Opisthobranchia, Anaspidea Aplysia californica - - AY569552

Opisthobranchia, Cephalaspidea Hydatina physis - - DQ991932

Opisthobranchia, Cephalaspidea Pupa strigosa - - AB028237

Opisthobranchia, Notaspidea Berthellina ilisima - - DQ991929

Opisthobranchia, Nudibranchia Chromodoris magnifica - - DQ991931

Opisthobranchia, Sacoglossa Ascobulla fragilis - - AY345022

Locality data and museum catalogue numbers of vouchers are indicated for the genomes newly sequenced here. Abbreviations for museum collections holdingthe vouchers are: California Academy of Sciences, San Francisco, United States of America (CASIZ); Natal Museum, Pietermaritzburg, South Africa (NM); FloridaMuseum of Natural History, University of Florida, Gainesville, Florida, United States of America (UF). An asterisk (*) next to a Genbank accession number indicatesthe 10 complete genome sequenced for the present study.

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 4 of 15

Table 2 Structural features of the 10 new mitochondrial genomes

Auriculinellabidentata

Myosotellamyosotis

Ovatellavulcani

Pedipespedipes

Peroniaperonii

Salinatorrhamphidia

Succineaputris

Rhopalocaulisgrandidieri

Trimusculusreticulatus

Siphonariagigas

Totalsize

14, 135 14, 215 14, 274 16, 708 13, 968 14, 007 14, 092 14, 523 14, 044 14, 518

%A 25.8 25.2 25.0 28.6 27.1 26.7 33.8 29.3 26.4 24.3

%T 30.9 32.5 29.7 33.7 37.3 35.6 42.9 33.9 34.7 37.2

%C 20.4 20.3 21.6 18.4 15.4 16.9 10.8 19.7 18.2 15.1

%G 22.6 22.0 23.7 19.3 20.3 20.8 12.1 17.1 20.6 23.4

%A+T 56.7 57.7 54.7 62.3 64.4 62.3 76.7 63.2 61.1 61.5

POR 44 46 44 ? (397) 54 44 47 42 47 72

rrnL 1, 037 1, 041 1, 057 1, 138 1, 033 1, 025 1, 020 1, 006 1, 057 1, 109

rrnS 704 714 711 786 714 714 755 716 719 755

cob 1, 111 (TTG/T) 1, 110 (TTG/TAG) 1, 110 (TTG/TAA) 1, 108 (TTG/T) 1, 108 (TTG/T) 1, 111 (TTG/T) 1, 107 (TTG/TAG) 1, 068 (TTG/TAG) 1, 110 (TTG/TAA) 1, 125 (TTG/TAA)

cox1 1, 533 (TTG/TAA) 1, 527 (ATG/TAA) 1, 533 (TTG/TAA) 1, 527 (TTG/TAA) 1, 525 (TTG/T) 1, 527 (TTG/TAA) 1, 548 (TTG/TAG) 1, 527 (TTG/TAA) 1, 527 (TTG/TAG) 1, 530 (GTG/TAG)

cox2 687 (TTG/TAG) 669 (GTG/TAA) 666 (TTG/TAG) 681 (GTG/TAA) 666 (TTG/TAA) 664 (GTG/T) 649 (ATG/T) 634 (TTG/T) 666 (TTG/TAA) 672 (GTG/TAA)

cox3 778 (ATG/T) 780 (ATG/TAG) 778 (ATG/T) 780 (GTG/TAG) 804 (ATG/T) 781 (ATG/T) 783 (ATG/TAG) 772 (ATG/T) 778 (ATG/T) 810 (ATG/TAA)

nad1 906 (GTG/TAA) 906 (TTG/TAG) 906 (TTG/TAA) 903 (ATG/TAG) 906 (TTG/TAA) 958 (TTG/T) 916 (TTG/T) 919 (TTG/T) 906 (TTG/TAA) 903 (TTG/TAA)

nad2 945 (ATG/TAG) 948 (ATG/TAG) 942 (CTG/T) 927 (ATG/TAA) 939 (GTG/TAG) 925 (TTG/T) 975 (TTG/TAA) 915 (GTG/TAA) 916 (TTG/T) 942 (ATG/TAA)

nad3 354 (TTG/TAA) 349 (TTG/TAA) 354 (TTG/TAA) 372 (ATG/TAA) 327 (ATT/TAA) 349 (TTG/T) 352 (ATG/T) 351 (GTG/TAA) 357 (ATG/TAG) 361 (ATG/T)

nad4 1, 311 (TTG/TAG) 1, 305 (TTG/TAA) 1, 308 (TTG/TAG) 1, 302 (ATG/TAG) 1, 318 (TTG/T) 1, 311 (TTG/TAG) 1, 326 (ATG/TAA) 1, 308 (ATG/TAA) 1, 306 (TTG/T) 1, 324 (TTG/T)

nad4L 327 (TTG/TAG) 291 (TTG/TAA) 286 (TTG/T) 288 (GTG/TAG) 283 (ATG/T) 279 (TTG/TAG) 275 (ATA/TA) 274 (TTG/T) 286 (ATG/T) 294 (GTG/TAA)

nad5 1, 665 (ATA/TAG) 1, 689 (ATA/TAG) 1, 671 (TTG/TAG) 1, 701 (ATG/TAG) 1, 671 (TTG/TAG) 1, 671 (TTG/TAG) 1, 680 (ATG/TAA) 1, 662 (ATT/TAA) 1, 680 (TTG/TAG) 1, 674 (GTG/TAA)

nad6 483 (ATT/TAA) 477 (TTG/TAA) 480 (ATT/TAG) 462 (TTG/TAA) 468 (ATT/TAA) 474 (TTG/TAA) 453 (ATG/TAA) 450 (ATG/TAG) 456 (TTG/TAG) 489 (ATT/TAG)

atp6 643 (GTG/T) 645 (ATG/TAG) 645 (GTG/TAG) 642 (ATG/TAG) 642 (TTG/TAA) 643 (ATG/T) 657 (ATG/TAA) 618 (ATG/TAA) 643 (ATG/T) 657 (ATG/TAA)

atp8 157 (GTG/T) 153 (ATG/TAG) 159 (ATG/TAG) 153 (ATG/TAA) 153 (ATG/TAA) 151 (ATG/T) 123 (TTG/TAA) 151 (GTG/T) 186 (ATG/TAG) 162 (ATG/TAA)

The size of each genome and the POR are in bp. Start and stop codons for protein-coding genes are indicated in parentheses.

White

etal.BM

CEvolutionary

Biology2011,11:295

http://www.biom

edcentral.com/1471-2148/11/295

Page5of

15

Pedipes pedipes. The typical size of the intergenic spacersvaries from 1 to 80 bp. A large intergenic spacer (651 bp)is found in Rhopalocaulis grandidieri between trnS1,trnS2, and another one (270 bp) in Ovatella vulcanibetween trnM, nad3. Of the 20 intergenic spacers foundin Pedipes pedipes, five of them are sizeable and AT-rich:288 bp between nad6, nad5, 447 bp between trnS2, trnT,397 bp between cox3, trnQ, 700 bp between trnR, trnS1,and 317 between nad4, trnI. An AT-rich intergenicspacer between cox3, trnI is found in all 10 mitochondrialgenomes, and this was determined to be the potential ori-gin of replication (POR) which concurs with previousfindings [6]. Due to gene rearrangements, the POR forPedipes pedipes may be located between cox3, trnQ (397bp) or between nad4, trnI (317 bp). The POR of Pedipespedipes is likely adjacent to the start of cox3 for two rea-sons. First, the percentage of A+T in the 50 bp adjacentto cox3 is higher than the 50 bp adjacent to trnI (62.0%and 58.8%, respectively). Second, a POR adjacent to cox3was previously found in other gastropod mitochondrialgenomes [10].

The 22 tRNAs for each of the 10 mitochondrial gen-omes were easily located, with the exception of trnS1 inRhopalocaulis grandidieri. Unlike the trnS1 of the othernine mitochondrial genomes, the anticodon loopsequence is CTGCTAG instead of the typical CTGCTAAand there are two base mispairings in the anticodon loop,which is also uncommon (region sequenced with a 4Xcoverage). All of the trnS1 and trnS2 genes lack the DHUstem.

Molecular phylogenyThe trees from the ML and Bayesian analyses shareidentical topologies and similar branch lengths and nodesupport values (Figure 3). Pulmonates form a paraphy-letic group at the base of the Euthyneura; Pyramidelladolabrata, traditionally regarded as a basal heterobranchexternal to Euthyneura, is nested within (paraphyletic)pulmonates. Siphonaria, traditionally regarded as a basalpulmonate, is nested within Opisthobranchia whichform a derived, monophyletic group at the tip of theeuthyneuran tree. The veronicellid slug Rhopalocaulis

Figure 3 Phylogenetic relationships within euthyneurans (pulmonates and opisthobranchs). Topology obtained from the ML and BIanalyses, based on the concatenated amino acid sequences of the 13 protein-coding genes. Opisthobranchs are indicated in green andpulmonates in blue. Node support values are cited using the following format: ‘’100/1’’ means that ML bootstrap value = 100%, and that BIposterior probability = 1.00. Branch lengths depicted here are from the Bayesian analysis (-LnL = -99753.12) performed with MTRev+I+G modelusing Mr Bayes in Topali. The ML bootstrap values indicated here are from the ML analysis performed with PhyML in Topali (MTRev+I+G, log-likelihood -99834.23). Alternative analyses provided similar topology, branch lengths, and support values.

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 6 of 15

grandidieri is the most basal species of the euthyneur-ans, with high node support. Freshwater snails Biompha-laria glabrata and Radix balthica form a clade emergingjust after Rhopalocaulis grandidieri. Land snails Succineaputris and Albinaria coerulea form a clade emerging justafter the freshwater snails. The ellobiids (Pedipes ped-ipes, Myosotella myosotis, Ovatella vulcani, and Auricu-linella bidentata) are not monophyletic and are spreadthroughout the basal portion of the tree. However, thefalse limpet Trimusculus reticulatus (Trimusculidae) isrecovered as closely related to two ellobiids (Ovatellavulcani and Auriculinella bidentata), which is stronglysupported. The three species of onchidiids (Peronia per-onii, Onchidella borealis, and Onchidella celtica) form ahighly-supported monophyletic group. Salinator rham-phidia (Amphiboloidea), traditionally regarded as a basalpulmonate (mainly because of the presence of an oper-culum), is not basal with respect to all pulmonates.Within the clade Opisthobranchia, Ascobulla fragilis isfound to be the most basal. AU and SH tests performedindicate that other alternative hypotheses of monophyly(Eupulmonata, Geophila, Ellobiidae, and Systellommato-phora) are rejected (Table 3).Some nodes are strongly supported, with Bayesian pos-

terior probabilities of 1 and bootstrap values of 100%,such as the monophyly of opisthobranchs (includingSiphonaria) and the monophyly of the clade includingTrimusculus and two ellobiids (Ovatella and Auriculi-nella). Some other nodes are still reasonably supported,although less strongly, with Bayesian posterior probabil-ities of 1 and bootstrap values superior to 75%. However,four of the deeper nodes display no statistical support,with Bayesian posterior probabilities less than 0.8 andbootstrap values less than to 50%. Those nodes, indicatedas thin lines in Figure 3, should be regarded as unre-solved polytomies, which directly affects the interpreta-tion of the relationships (see Discussion).

Genome organization and rearrangementsBiomphalaria glabrata, Salinator rhamphidia, Trimus-culus reticulatus, Ovatella vulcani, Auriculinella biden-tata, Peronia peronii, Onchidella borealis, Onchidellaceltica share an identical mitochondrial genome organi-zation (Figure 4). Albinaria coerulea shares the samegenome organization except that the position of trnS1and trnS2 are switched and trnS1 is inverted. The gen-ome of the two freshwater snails Radix balthica andBiomphalaria glabrata differ in the location of fivetRNAs (P, H, G, C, and Y). The genome of Rhopalocau-lis grandidieri differs in the location of seven tRNAs (C,F, G, W, H, L2, and E). The genome of Succinea putrisdiffers in the location of three tRNAs (F, Y, and W),with the latter two genes being coded on the minusstrand instead of the plus strand. In Pedipes pedipes,trnT, cox3 swapped with trnS1, nad4, and trnQ andtrnR moved between these two swapped sets of genes.Myosotella myosotis only differs from the most standardmitochondrial genome organization by the rearrange-ment of nad4L between cox2 and trnY. The location oftrnY before cox1 is a unique and unusual feature of themitochondrial genome of Pyramidella dolabrata. Otherattributes, such as the location of atp6 prior to atp8 andthe encoding of trnG by the minus strand, are exclusiveto Pyramidella dolabrata.The opisthobranch genomes only differ from the com-

mon mitochondrial genome arrangement in the positionof trnY, trnW, and trnC. The trnC is located betweentrnH and trnQ in Ascobulla fragilis, Chromodoris magni-fica, and Berthellina ilisima, and between trnN and atp6in Aplysia californica, Pupa strigosa, and Hydatina phy-sis. The organization of the genome of Siphonaria gigasis more similar to the genome of opisthobranchs thanpulmonates, which is also true for Siphonaria pectinatadespite some moderate rearrangements (e.g., trnY andtrnW are adjacent to nad4L).

Table 3 Statistical tests of alternative phylogenetic relationships

Topology * Log-likelihood AU SH

Unconstrained: (Rh, ((Ra, Bi), ((Su, Al), (Ped, ((Mm, Mm), (Pyr, ((Sal, ((Tr,(Ov, Au)), (Per, (Ob, Oc)))), (As, ((Sp, Sg), (Ap, ((Pu, Hy)(Ch, Be))))))))))))

-99834.23 0.99 0.99

Eupulmonata: (Pyr, (As, ((S.p, S.g), (Ap, ((Pu, Hy)(Ch, Be))))), (Sal, (Tr, ((Ra, Bi),(Rh,, ((Su, Al), (Ped, ((Mm, Mm), ((Ov, Au), (Per, (Ob, Oc)))))))))))

-99872.19 < 0.01 < 0.01

Geophila: (Pyr, (As, ((Sp, Sg), (Ap, ((Pu, Hy)(Ch, Be))))), (Sal, (Tr, ((Ra, Bi), ((Ped,((Mm, Mm), ((Ov, Au), (Rh, ((Su, Al), (Per, (Ob, Oc))))))))))))

-99929.87 < 0.01 0.01

Ellobiidae: (Pyr, (As, ((Sp, Sg), (Ap, ((Pu, Hy)(Ch, Be))))), (Sal, (Tr, ((Ra, Bi),((Ped, ((Mm, Mm), (Ov, Au))), (Rh, ((Su, Al), (Per, (Ob, Oc)))))))))

-99956.03 < 0.01 0.01

Systellommatophora: (Pyr, (As, ((Sp, Sg), (Ap, ((Pu, Hy)(Ch, Be))))), (Sal, (Tr,((Ra, Bi), ((Ped, ((Mm, Mm), (Ov, Au))), (((Per, (Ob, Oc)), Rh), (Su, Al)))))))

-999997.54 < 0.01 < 0.01

*: Rh = Rhopalocaulis grandidieri; Ra = Radix balthica; Bi = Biomphalaria glabrata; Su = Succinea putris; Al = Albinaria coerulea; Ped = Pedipes pedipes; Mm =Myosotella myosotis; Pyr = Pyramidella dolabrata; Sal = Salinator rhampidia; Tr = Trimusculus reticulatus; Ov = Ovatella vulcani; Au = Auriculinella bidentata; Per =Peronia peronii; Ob = Onchidella borealis; Oc = Onchidella celtica; As = Ascobulla fragilis; Sg = Siphonaria gigas; Sp = Siphonaria pectinata; Ap = Aplysia californica;Pu = Pupa strigosa; Hy = Hydatina putris; Ch = Chromodoris magnifica; Be = Berthellina ilisima.

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 7 of 15

DiscussionSpecies diversity of the pulmonate gastropods is largelydominated by the land snails and slugs, or Stylommato-phora, which include at least 25, 000 species. However,less than five percent of the pulmonate species diversity isfound in the ten other higher taxa, which are anatomicallyand ecologically very distinct from each other: Amphibo-loidea, Ellobiidae, Hygrophila, Onchidiidae, Otinidae,Siphonariidae, Smeagolidae, Trimusculidae, Veronicelli-dae, and Williamiidae. The present contribution constitu-tes a significant increase in taxon sampling for completemitochondrial genomes of pulmonates, especially regard-ing the non-stylommatophoran taxa (Figure 2, Table 1):complete mitochondrial genomes of amphiboloids, trimus-culids, veronicellids, and two ellobiid “subfamilies” (Pedi-pedinae and Ellobiinae) are presented here for the firsttime. As of today, complete mitochondrial genomes areunavailable for only three of the pulmonate higher taxa:Otinidae (only one species known), Williamiidae (less than10 species), and Smeagolidae (less than 10 species).

The topology of the consensus tree produced by ourphylogenetic analyses (Figure 3) is very similar to the treesobtained previously based on all protein-coding genes ofthe mitochondrial genomes [6,7]: pulmonates are paraphy-letic, at the base of euthyneurans (opisthobranchs and pul-monates); opisthobranchs (including Siphonaria) aremonophyletic, at the tip of the tree. Pyramidella dolab-rata, traditionally regarded as a basal heterobranch (out-side euthyneurans), is more closely related to pulmonates.Biomphalaria glabrata is found here to be at the base ofpulmonates [7] instead of at a more derived position inthe tree [6].The trees based on mitochondrial genomes (Figure 3)

show some congruence with the trees based on individualgenes -18S, 28S, 16S, COI (Figure 1). In particular, bothdata sets agree on some nodes that were unsuspectedbefore, such as the close relationship of Trimusculus reti-culatus and ellobiids. However, both data sets also showsome noticeable differences. This might be explained bythe fact that some of the deeper nodes are poorly

Figure 4 Hypothesized gene rearrangements of the genomes included in the present study. The phylogenetic topology is from thepresent study. Genes encoded by the minus strand are underlined.

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 8 of 15

supported by the mitochondrial AA data set presentedhere (Figure 3). Also, the taxon sampling differs greatly:nearly 80 species of pulmonates are included in a recentanalysis using 18S, 16S, and COI sequences [5] whileonly 20 complete mitochondrial genomes of pulmonatesare available, including those generated in this study.Below we discuss the impact of our phylogenetic resultson our understanding of the evolution of pulmonate gas-tropods. In particular, we focus on the new insights andnew questions raised by the addition of the 10 new mito-chondrial genomes.In the present results, the most basal lineage of all

euthyneurans are the terrestrial, veronicellid slugs, repre-sented here by one species, Rhopalocaulis grandidieri(Figure 3). Traditionally, Veronicellidae has been classifiedin Systellommatophora, along with Onchidiidae andRathouisiidae [12], although no synapomorphies could befound for systellomatophorans in cladistic analyses [1].The representation of veronicellids in molecular analysesis recent, but trees based on individual gene sequencessupport the monophyly of Systellommatophora (Figure 1)[4,5]. A basal position of veronicellids with respect toother pulmonates was proposed based on morphology[13,14]. However, this result was based on problematicinterpretations of the anatomy of systellommatophorans,such as the fact that they supposedly lack a pneumostomeand a lung. If confirmed, a basal position of veronicellidscould lead us to re-interpret their extremely reduced pal-lial cavity. However, additional mitochondrial genomes ofveronicellids will be necessary for testing this new result,because a single genome might introduce a bias (e.g.,long-branch attraction) in the analyses.Freshwater snails (here represented by Radix balthica

and Biomphalaria glabrata) are found to be among themost basal lineages, although their relationships withrespect to Stylommatophora (here represented by Succineaputris and Albinaria coerulea), are unclear. However, theposition of Hygrophila is still quite unstable and does varyfrom one study to another [2,4,5]. For unclear reasons, ourattempts to get new complete genomes of freshwater pul-monates failed (very few long PCRs worked), although westarted with fresh material from 13 species).A relatively basal position of Stylommatophora with

respect to other pulmonates was obtained in previous ana-lyses based on mitochondrial genomes and individualgenes (Figure 1) [5-7]. Many derived features characterizeland snails and slugs, especially in the excretory and pul-monary systems, which are physiologically critical for lifeon land [1,12]. As a result, land snails and slugs have oftenbeen regarded as the most derived -and thus most recent -lineage of pulmonates. However, several of those derivedfeatures are autapomorphic (e.g., ureter anatomy) orpotentially homoplasic (e.g., eyes at the tip of the eyetentacles) [1]. A basal position of land snails and slugs

supports Solem’s theory according to which stylommato-phorans were the first pulmonates to emerge 350 Mya[5,12]. Although stylommatophorans are among the mostbasal pulmonates and euthyneurans here, their relation-ships with respect to freshwater snails (Hygrophila) areunclear because of low node support.The Ellobiidae is a diverse taxon with about 800 species

names available in the literature, although only 250 ofthem are likely to be valid. More importantly, Ellobiidae isphylogenetically diverse because the 24 ellobiid genera arecharacterized by different combinations of plesiomorphicand derived characters [15,16]. As a result, no exclusivemorphological synapomorphy can be found for ellobiidswithin the broader context of all pulmonates [1]. Formany years, molecular analyses could not adequately testthe phylogenetic status of ellobiids because of low taxonsampling [2,6]. However, ellobiids were found to be mono-phyletic in a recent molecular analysis based on a new,comprehensive data set including 25 ellobiid species(Figure 1) [5], suggesting that additional sequenced mito-chondrial genomes might be needed to properly addressthe phylogenetic status of the Ellobiidae, especially consid-ering that two subfamilies (Carychiinae and Melampodi-nae) are still not represented.Although the phylogenetic status of Ellobiidae remains

problematic, analyses based on concatenated mitochon-drial AA sequences and those based on individual genes(18S, 28S, 16S, COI) all agree that the false limpet Trimus-culus reticulatus is closely related to ellobiids (Figure 1)[2-5]. Trimusculus reticulatus also shares the same gen-ome organization as several ellobiid species (Figure 4).This is a new result that had not been suggested usingmorphology. In fact, there is not any obvious anatomicalfeature that appears to be shared by Trimusculus and ello-biids [5]. Some features are found in ellobiids and Trimus-culus, but they are also found in many other pulmonates,such as the globineurons and a hypoathroid central ner-vous system [1,17]. Another interesting result obtainedhere, as well as in recent analyses based on individualgenes, is the close relationship between Onchidiidae andsome ellobiids (Figure 1). Onchidiidae and Ellobiidae weresuggested to be classified together in the Ellobioidea [18],partly based on characters from the nervous system.Our present data (Table 3) reject the Geophila

hypothesis (Systellommatophora and Stylommatophora)which is also not supported by other recent moleculardata (Figure 1). Morphological data supported themonophyly of Geophila, mainly based on two synapo-morphies, i.e., the loss of heterostrophy and the pre-sence of eyes at the tip of the ocular tentacles [1]. Bothof these features might have evolved independently inStylommatophora, Onchidiidae, and Veronicellidae. TheEupulmonata hypothesis sensu Morton (Geophila andEllobiidae) also is rejected here [15].

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 9 of 15

The present analyses confirm two important resultsfrom recent molecular studies (Figure 1) [2-6]. First,Amphiboloidea, traditionally regarded as one of the mostbasal lineages of pulmonates, is not basal, suggesting thattheir operculum (amphiboloids and glacidorbids are theonly pulmonates with an operculum) was acquired secon-darily. Second, the pyramidelloids, traditionally regardedas basal (i.e., non-euthyneuran) heterobranchs, are nowconsistently found to be nested within pulmonates[2-5,19]. However, the pyramidelloids have always beendifficult to classify. Some past authors have even regardedthem as opisthobranchs [20-23]. The fact that pyramidel-lids differ so much from other pulmonates may be relatedto the fact that they live submerged in the seawater(Williamia species are the only other pulmonates that livesubmerged as adults).Finally, the present data confirm that Siphonaria is

more closely related to opisthobranchs than to pulmo-nates, although this relationship is much clearer withmitochondrial genomes than with individual genes (18S,28S, 16S, and CO1) [2-6,8]. More specifically, Siphonariaappears to be most closely related to sacoglossans (hererepresented as Ascobulla fragilis). For most of the 20th

century, authors have classified Siphonaria withinPulmonata, mainly because they tend to live in the upperintertidal zone, exposed to the air for most of the day,unlike opisthobranchs, which all live submerged evenwhen found in the intertidal zone. As a result, the pallialcavity of Siphonaria has been interpreted as a pulmonarycavity and their gills as secondary gills. However, earlyanatomists had recognized the similarity of the pallial gillof Siphonaria with that of cephalaspideans [24,25], sug-gesting that they could be homologous [1,26]. The pallialgill of Siphonaria was even sometimes referred to as acephalaspidean gill [27]. Siphonaria and opisthobranchsshare another important feature, i.e., the production of amilky white substance (polypropionate metabolites) whenirritated [28], although defensive metabolites are alsofound in other pulmonates, such as Onchidium (Onchi-diidae) and Trimusculus [29]. Even the gene order of themitochondrial genome of Siphonaria gigas (see the posi-tion of trnY, trnW, nad4L, and cob; Figure 4) supports anaffinity with opisthobranchs.The order of the protein-coding, rRNA, and tRNA genes

in mitochondrial genomes is known to be potentiallyinformative for phylogenetics [30]. However, the genomesof pulmonate gastropods display limited informative varia-tion because the organization of protein-coding and ribo-somal genes is stable across euthyneurans (Figure 4). Also,Figure 3, which places gene order in a phylogenetic con-text, suggests that species may share the same gene orderbecause of convergence. Indeed, the freshwater snail Biom-phalaria glabrata shares the same gene order as other dis-tantly-related pulmonates represented here (Salinator,

Trimusculus, Ovatella, Auriculinella, Onchidella, andPeronia), but differs from the other freshwater snail repre-sented here, Radix balthica, for five tRNA genes. So, thegene order found in Biomphalaria glabrata and other pul-monates likely is due to convergence, i.e., tRNA genesended in identical positions through distinct but conver-gent series of rearrangements. The fact that the genomesof the two Siphonaria species differ in the position of twoprotein-coding genes (nad4 and nad4L), which has beenexceptionally observed in euthyneurans, supports thesame idea that the gene order of the mitochondrial gen-ome of a given species should probably not be used toextrapolate on the gene order of a whole lineage. In caeno-gastropods, Rawlings and collaborators also demonstratedthat two closely-related species of Dendropoma can differsignificantly in genome organization [11]. When additionalgenomes are available for each major lineage of pulmonate(Veronicellidae, Hygrophila, Onchidiidae, etc.), it mightbecome possible to determine a common gene order toeach lineage, but it is not possible at this stage because ofour still limited number of complete genomes. There is lit-tle doubt, however, that, as additional complete genomesbecome available in euthyneurans, we will discover anincreased variability in euthyneuran gene order.

ConclusionThe present data constitute a significant increase intaxon sampling for complete mitochondrial genomes ofpulmonates (Figure 2). However, despite these efforts, weare still far from a comprehensive understanding ofhigher relationships of pulmonates. Most of the deepnodes are still uncertain, mainly due to low supportvalues as well as some incongruence between analysesbased on different data sets (complete mitochondrialgenomes versus individual genes -18S, 28S, 16S, CO1),these two issues being obviously related (Figures 1, 3).Additional pulmonate genomes are needed, especially forthe taxa for which no genome is currently available(Williamia, Otina, and Smeagol). Euthyneuran relation-ships would also greatly benefit from the addition of newgenomes of basal heterobranchs (Architectonicoidea,Valvatoidea, Omalogyroidea, Rissoelloidea, Orbitestelli-dae). Being the most closely related taxa to euthyneurans,basal heterobranchs could help stabilize the topologywithin euthyneurans.Euthyneuran phylogenetics and, more broadly, mollus-

can phylogenetics are still based on limited data, at leastcompared with vertebrate or arthropod phylogenetic stu-dies which may include more than 40 kb of sequence data.Addressing some of the incongruence between trees basedon individual genes (18S, 28S, 16S, CO1) and those basedon complete mitochondrial genomes in euthyneuran phy-logenetics is a long-term goal that is beyond the scopeof the present study. Progress will obviously require

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 10 of 15

increased taxon sampling, but, more importantly, newmarkers, especially given that, in some ways, using mito-chondrial genomes is similar to analyzing one marker withmany loci [5]. In the future, when the number and diver-sity of molecular sequences and taxon samples areincreased, merging all data into a single data set may alsohelp provide greater resolution.Meanwhile, we have to accept that our progress regard-

ing euthyneuran phylogenetics will be slow. Step by step,important discoveries are being made (Figures 1 and 3).Some of the results obtained from recent molecular studiesconstitute new, major findings [2-5]. In particular, all datasets support close relationships between the false limpetTrimusculus reticulatus and ellobiids. Also, pyramidelloids,regarded for many decades as basal (non-euthyneuran) het-erobranchs, are nested within pulmonates. Finally, the falselimpets Siphonaria, traditionally regarded as pulmonates,are closely related to the sacoglossan opisthobranchs; mito-chondrial genomes strongly suggest that Siphonariais nested within the monophyletic Opisthobranchia (Figure3), while individual genes suggest that Siphonaria andsacoglossans could be at the base of the monophyletic Pul-monata (Figure 1). All those results reinforce the idea thatopisthobranch and pulmonate phylogenetics and evolutionshould be studied together, as euthyneurans [1,26].

MethodsTaxon SamplingIn addition to the ten mitochondrial genomes success-fully sequenced for the present study, our analyses alsoinclude 17 complete mitochondrial genomes of gastro-pods obtained from GenBank (Table 1). Of the ten mito-chondrial genomes of pulmonates available (Figure 2),three were not selected here: the sequence of Cepaeanemoralis, generated more than 15 years ago, is of poorquality; the mitochondrial genomes of Biomphalariatenagophila and Platevindex mortoni, although publiclyavailable in Genbank, are still unpublished and should bepublished shortly by their authors. In addition, sevengenomes were selected to represent the major lineages ofopisthobranchs (Anaspidea, Cephalaspidea, Notaspidea,Nudibranchia, and Sacoglossa). Three outgroups wereselected within Caenogastropoda.

Species identification, Vouchers, and DNA extractionSpecies for which genomes were obtained in the presentstudy were identified by taxonomic experts (some ofwhom are co-authors of the present article): RosemaryGolding identified the Salinator, Suzete Gomes Rhopalo-caulis, Antonio M. de Frias Martins the ellobiids, TracyWhite the Siphonaria, and Benoît Dayrat the Peronia,Succinea, and Trimusculus. Voucher specimens aredeposited in museum collections (Table 1). For each spe-cies, the complete mitochondrial genome was obtained

from a single individual (with the exception of the tinyPedipes pedipes for which three individuals had to beused). In most cases, that individual is part of the lotdeposited as voucher. However, the small specimen usedfor DNA extraction of Salinator, Auriculinella, and Ova-tella had to be destroyed. In these cases, the voucher lotcontains other individuals from the same population.DNA was extracted using the phenol-chloroform extrac-tion protocol with cetyltrimethyl-ammonium bromide(CTAB) [31].

PCR Amplification and SequencingThree approaches were combined to successfully obtaincomplete mitochondrial genomes (Additional file 1): 1)long PCR products (from ~3 to ~10 kb) were amplifiedand sequenced using shotgun sequencing; 2) short PCRproducts (less than ~1.5 kb) were amplified using pulmo-nate-specific primers spanning the entire genome and spe-cifically designed for the present study (pulmonate-specificprimers were designed through the alignment of all thesequences of pulmonate mitochondrial genomes availablewhen the present study started); 3) short PCR productswere amplified by ‘primer-walking, ‘ i.e., using individual-specific primers designed in previously sequenced regions(by individual-specific primer, we mean that distinct pri-mers were designed for each genome being sequenced).Sequence coverage ranged from 2X (sequences from pul-monate- and individual-specific primers) to greater than20X (sequences from shotgun sequencing of long PCRs).Only high-quality chromatograms were utilized in areaswhere sequence coverage was 2X. Additional file 1 sum-marizes how these three approaches were combined toobtain each genome.Long PCR amplificationFive pairs of universal primers were utilized to generateshort PCR products within five mitochondrial genes: cox1,cox3, cob, rrnS, and rrnL [32]. In order to increase the suc-cess of long PCR amplification, individual-specific primerswere designed within the short PCR fragments of cox1,cox3, cob, rrnS, and rrnL obtained with universal primers.Different combinations of individual-specific primers werethen used to amplify large regions of the genome (from ~3to ~10 kb). In some cases, universal primers were com-bined with individual-specific primers (combinations oftwo universal primers very rarely yielded successful ampli-fications). The 25 μl short PCR reactions contained 10.9 μlof water, 2.5 μl of 10X PCR Buffer, 2 μl of 25 mM MgCl2,1 μl of each 10 μM primer, 2 μl of dNTP Mixture, 0.2 μl(1 unit) of TaKaRa Taq (Code No. R001A), 5 μl of 20 ng/μl template DNA, and 0.4 μl of 100X BSA (10 mg/mlBovine Serum Albumin). In some short PCR reactions,BSA was replaced with 5 μl of 5X Qiagen Q-Solution(water added to these reactions was 6.3 μl). The thermo-profile used for cox3, cob, and rrnS was five minutes at

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 11 of 15

94°C; 40 cycles of 40 seconds at 94°C, 1 minute at 46°C,and 1 minute at 72°C; and 10 minutes at 72°C. The ther-moprofile used for cox1 and rrnL was five minutes at94°C; 30 cycles of 40 seconds at 94°C, 1 minute at 46°C,and 1 minute at 72°C; and 10 minutes at 72°C. The 25 μllong PCR reactions contained 3.8 μl of water, 2.5 μl of10X LA PCR Buffer II, 0.5 μl of 25 mM MgCl2, 5 μl ofeach 2 μM primer, 3 μl of dNTP Mixture, 0.2 μl (1 unit)of TaKaRa LA Taq (Code No. RR002M), and 5 μl of20 ng/μl template DNA. The thermocycler profile for thelong PCRs consisted of one minute at 98°C, followed by30 cycles of 98°C for 10 seconds and 68°C for 15 minutes,with a final extension of 10 minutes at 72°C. All long andshort PCR products were cleaned with Qiagen QIAquickPCR Purification Kit (Cat. No. 28106) prior to sequencing.Shotgun sequencing of long PCR productsThe purified long PCR products were sheared using theHydroShear from GeneMachines. The sheared DNA wasthen blunt-end repaired and visualized on a 1% agarosegel stained with ethidium bromide. Gel fragments of thecorrect size (~250bp) were excised and purified from thegel. The repaired DNA fragments were then ligated intothe pmcl vector plasmid and transformed into competentEscherichia coli cells by electroporation. The cells werethen blue/white screened on agar plates and 96 whitecolonies were picked to form each clone library. A ran-dom selection of the library was then amplified by PCRto ensure the DNA insert was present. The libraries weresequenced at the Joint Genome Institute (JGI), WalnutCreek, California, in the context of a Genomics coursetaught by Dr. Mónica Medina, in collaboration betweenthe University of California at Merced and the JGI.Short (pulmonate-specific) PCR amplificationOnly one genome was sequenced in its entirety using onlylong PCR and shotgun sequencing (Trimusculus reticula-tus; see Additional file 1). For six other genomes, longPCR and shotgun sequencing yielded partial genomes withgaps of various sizes that needed to be closed. A set of 94pulmonate-specific primers were designed (Additionalfile 2) to close those gaps in the six partial mitochondrialgenomes obtained through shotgun sequencing. Multiplecombinations of pulmonate-specific primers were used toamplify gaps in partial genomes obtained from shotgunsequencing. Those combinations were also used to amplifygenomes of species without going through shotgunsequencing (see, Additional file 1). In some cases, depend-ing on the strand used for transcription, a combination oftwo forward primers (e.g. Nad4-879F and Cox3-164F) ortwo reverse primers (e.g. Nad3-128R and Nad4-642R) hadto be used. For some individuals (e.g. Salinator rhamphi-dia and Rhopalocaulis grandidieri), the long PCR productwas used as the template for amplification instead of thegenomic DNA (for increasing chances of successful ampli-fication). For those PCRs targeting gaps of ~900 bp or

greater, the long PCR reaction and thermoprofile was used(see above). Short PCR reactions (see above) were used forgaps less than 900 bp with the following thermoprofile:two minutes at 94°C; 5 cycles of 40 seconds at 94°C, 45seconds at 45°C, and 1 minute at 72°C; 30 cycles of 40 sec-onds at 94°C, 40 seconds at 55°C, and 1 minute at 72°C;and 3 minutes at 72°C. PCR products were cleaned witheither the Qiagen QIAquick PCR Purification Kit or Exo-SAP (2 μl of 1u/μl Shrimp Alkaline Phophatase, 0.1 μl of20u/μl Exonuclease I, and 6 μl of water per 4 μl of PCRproduct) and sent for sequencing. Some PCR productsproved difficult to sequence directly (typically longer PCRproducts) and were cloned using the Promega pGEM-TEasy Vector System II (Cat. No. A1380) then cleaned withthe Promega Wizard Plus SV Minipreps DNA purificationSystem (Cat. No. A1330) prior to sequencing (productswere sent out for sequencing).Short (individual-specific) PCR amplificationFor the gaps that remained in the mitochondrial genomesafter using the pulmonate-specific primers and shotgunsequencing, individual-specific primers were designedand the gaps were primer-walked until completion of thegenome (Additional file 1). In some instances, a combina-tion of a pulmonate-specific and an individual-specificprimer was used in an attempt to close gaps. PCR condi-tions for the individual-specific primers were identical tothose used for the pulmonate-specific primers.

Genome assembly and annotationThe shotgun sequence chromatograms produced by theJGI were read, bases were called, and a value wasassigned to the quality of the called bases by the Phredprogram [33]. The individual sequences were assembledinto contigs using Phrap http://www.phrap.org. Thecontigs created by Phrap were analyzed and assembledto form longer, more complete contigs using Consed[34]. Contig sequences from the shotgun sequencingwere saved in MacVector with Assembler 9.5.2 http://www.macvector.com. All other sequences (i.e., allsequences not obtained through sequencing at the JGI;see Additional file 1) were assembled in MacVector.Open reading frames (ORFs) of the assembled contigs

were analyzed by MacVector and the tentative identity ofeach protein-coding gene was determined based on themap of the mitochondrial genome of Aplysia dactylomela.Each gene was added to its corresponding Se-Al v2.0a11http://evolve.zoo.ox.ac.uk file containing the alignment ofmultiple homologous gastropods sequences and the genewas demarcated on the genome based on the results of thealignment. The limits of both the protein-coding andrRNA genes were adjusted manually based on location ofadjacent genes. All tRNA genes were located by handbased on the anticodon and the fairly conserved anticodonstem and loop sequence.

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 12 of 15

Phylogenetic analysesThe 13 protein-coding gene sequences were first trans-lated into amino acid sequences and then individuallyaligned using the default parameters of ClustalW (1.6) inMEGA version 4.1 (Beta) [35]. Each alignment wascropped to remove variation on either end and thensequences were concatenated. The concatenated protein-coding genes alignment (3, 458 amino acids) was adjustedmanually and a minimal number of sites were thenremoved (gaps created by insertions in the sequences ofthe caenogastropod outgroups). Three caenogastropodspecies were used as outgroups: Cymatium parthenopeum,Ilyanassa obsoleta, and Lophiotoma cerithiformis (Table 1).Because genes of animal mitochondrial genomes are fast

evolving [36-38], phylogenetic analyses based on aminoacid sequences are commonly preferred to analyses basedon nucleotide sequences [6,7,39]. Also, it has been shownthat individual genes provide less topological resolutionthan concatenated genes [10].The amino acid substitution model was determined to

be MTRev+I+G using the Akaike Information Criterion(AIC) in Topali version 2.5 [40]. The same model wasobtained whether each gene was treated independently asa distinct partition, or all genes were concatenated in asingle data set. Maximum likelihood analysis of the aminoacid was run using the MTRev+I+G model with PhyML inTopali, with bootstrap support values based on 1000 repli-cates. An alternative ML analysis was also run in RAxML[41] using MTzoa [42], a model specifically designed forlophotrochozoan mitochondrial data sets. Bayesian Metro-polis-coupled Markov chain Monte Carlo (MCMC) analy-sis was performed on the amino acid sequences using theMTRev+I+G model with MrBayes in the Topali interface(two parallel analyses, 100 million generations, sampledevery 100 generations, 25% burn-in). An alternative Baye-sian analysis was performed using the MTzoa model. TheTracer v1.4 [43] graphical tool was used to visualize con-vergence of the chains from Bayesian analyses. In addition,posterior samples from different (five) runs were com-pared, which is arguably the best approach to detectpotential problems of chain convergence [44].Alternative phylogenetic relationships were tested with

the Approximately Unbiased (AU) and Shimodaira-Hasegawa (SH) tests, using CONSEL [45] using defaultsettings. Alternative relationships were based on tradi-tional, morphology-based results as well as recent mole-cular results [1,5,26].

Additional material

Additional file 1: Approaches used to obtain the 10 newmitochondrial genomes in the present study. Three approaches werecombined to get each of the 10 mitochondrial genomes of the presentstudy: 1) shotgun sequencing; 2) sequencing of short and long PCR

fragments obtained with pulmonate-specific primers (designed for thepresent study; see Additional file 2); 3) sequencing of short and long PCRfragments obtained with individual-specific primers (designed for thepresent study). The asterisk (*) indicates individuals in which the longPCR product was used as the template for the majority of PCR reactionsusing pulmonate-specific and individual-specific primers.

Additional file 2: Pulmonate-specific primers designed for thepresent study. Those pulmonate-specific primers were designed: 1) bybuilding alignments for the sequences of each individual gene (exceptfor tRNAs) from the complete, pulmonate, mitochondrial genomesavailable prior to the present study (Figure 2, Table 1), and 2) by locatingconserved regions. All primers were specifically designed for the presentstudy, with the exception of ten of them: in cox1, F14 and R698 [59]; inrrnL, F437 and R972 [60]; in cob, F384 and R827 [61]; in rrnS, F302 andR695 [60]; in cox3, F174 and R713 [61].

AcknowledgementsLaboratory work for the present study was performed using funds from a USNational Science Foundation Grant (DEB-0933276, to B. Dayrat). Partialshotgun sequencing was donated by Eddy Rubin, director of the DOE JointGenome Institute, Walnut Creek, California, through the sequencing quotaallocated to the UC Merced Genome Biology course (BIO 142) taught byDr. Mónica Medina in Spring 2007 and Spring 2008. Bioinformatics analyses(assembly and annotation) were initiated by undergraduate and graduatestudents in the context of that course. Jennifer Kuehl provided technicalexpertise with the libraries at the JGI. Suzete Gomes provided tissue forRhopalocaulis grandidieri.

Author details1School of Natural Sciences, University of California, 5200 North Lake Road,Merced, CA 95343, USA. 2Australian Museum, 6 College Street Sydney NSW2010 Australia. 3CIBIO-Açores, Center for Biodiversity and Genetic Resources,Department of Biology, University of the Azores, 9501-801 Ponta Delgada,São Miguel, Azores, Portugal.

Authors’ contributionsTW participated in the molecular work and genome annotation, performedphylogenetic analyses, and wrote a first draft of the manuscript. MC and SBparticipated in the molecular work and RT in the genome annotation. RGand AFM provided fresh samples and commented on the manuscript. RTand SB participated in the research while undergraduates at the Universityof California at Merced. BD designed and developed the entire project. Allauthors read and approved the final manuscript.

Competing interestsThe authors declare that they have no competing interests.

Received: 6 May 2011 Accepted: 10 October 2011Published: 10 October 2011

References1. Dayrat B, Tillier S: Evolutionary relationships of euthyneuran gastropods

(Mollusca): a cladistic re-evaluation of morphological characters. Zool JLinn Soc 2002, 135:403-470.

2. Klussmann-Kolb A, Dinapoli A, Kuhn K, Streit B, Albrecht C: From sea toland and beyond - New insights into the evolution of euthyneuranGastropoda (Mollusca). BMC Evol Biol 2008, 8:57.

3. Dinapoli A, Klussmann-Kolb A: The long way to diversity - Phylogeny andevolution of the Heterobranchia (Mollusca: Gastropoda). Mol PhylogenetEvol 2010, 55:60-75.

4. Jörger KM, Stöger I, Kano Y, Fukuda H, Knebelsberger T, Schrödl M: On theorigin of Acochlidia and other enigmatic euthyneuran gastropods, withimplications for the systematics of Heterobranchia. BMC Evol Biol 2010,10:323.

5. Dayrat B, Conrad M, Balayan S, White TR, Albrecht C, Golding R, Gomes S,Harasewych MG, de Frias Martins AM: Phylogenetic relationships andevolution of pulmonate gastropods (Mollusca): new insights fromincreased taxon sampling. Mol Phylogenet Evol 2011, 59:425-437.

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 13 of 15

6. Grande C, Templado J, Zardoya R: Evolution of gastropod mitochondrialgenome arrangements. BMC Evol Biol 2008, 8:61.

7. Castro LR, Colgan DJ: The phylogenetic position of Neritimorpha basedon the mitochondrial genome of Nerita melanotragus (Mollusca:Gastropoda). Mol Phylogenet Evol 2010, 57:918-923.

8. Medina M, Lal S, Vallès Y, Takaoka T, Dayrat B, Boore J, Gosliner T: Crawlingthrough time: Transition of snails to slugs dating back to the Paleozoic,based on mitochondrial phylogenomics. Mar Genomics 2011.

9. Hatzoglou E, Rodakis GC, Lecanidou R: Complete sequence and geneorganization of the mitochondrial genome of the land snail Albinariacoerulea. Genetics 1995, 140:1353-1366.

10. Cunha RL, Grande C, Zardoya R: Neogastropod phylogenetic relationshipsbased on entire mitochondrial genomes. BMC Evol Biol 2009, 9:210.

11. Rawlings TA, MacInnis MJ, Bieler R, Boore JL, Collins TM: Sessile snails,dynamic genomes: gene rearrangements within the mitochondrialgenome of a family of caenogastropod molluscs. BMC Genomics 2010,11:440.

12. Solem A: Classification of the land Mollusca. In Pulmonates. Systematics,Evolution and Ecology. Volume 2A. Edited by: Fretter V, Peake J. New York:Academic Press; 1978:49-97.

13. von Salvini-Plawen L: The status of the Rhodopidae (Gastropoda:Euthyneura). Malacologia 1991, 32:301-311.

14. von Salvini-Plawen L, Steiner G: Synapomorphies and plesiomorphies inhigher classification of Mollusca. In Origin and evolutionary radiation of theMollusca. Edited by: Taylor J. London: The Malacological Society of London;1996:29-51.

15. Morton JE: The evolution of the Ellobiidae with a discussion on theorigin of the Pulmonata. Proc Zool Soc Lond 1955, 125:127-168.

16. de Frias Martins AM: Morphological and anatomical diversity within theEllobiidae (Gastropoda, Pulmonata, Archaeopulmonata). Vita Malacologica2007, 4:1-28.

17. Haszprunar G, Huber G: On the central nervous system of Smeagolidaeand Rhodopidae, two families questionably allies with theGymnomorpha (Gastropoda: Euthyneura). J Zool 1990, 220:185-199.

18. Tillier S: Relationships of Gymnomorph Gastropods Mollusca Gastropoda.Zool J Linn Soc 1984, 82:345-362.

19. Dinapoli A, Zinssmeister C, Klussmann-Kolb A: New insights into thephylogeny of the Pyramidellidae (Gastropoda). J Moll Stud 2011, 77:1-7.

20. Hyman LH: The Invertebrates: Mollusca I New York: McGraw-Hill; 1967.21. Thompson TE: In Biology of Opisthobranch Molluscs. Volume I. London: Ray

Society; 1976.22. Haszprunar G: On the origin and evolution of major gastropod groups,

with special reference to the Streptoneura. J Moll Stud 1988, 54:367-441.23. Wise JB: Morphology and phylogenetic relationships of certain

pyramidellids taxa (Heterobranchia). Malacologia 1996, 37(2):443-511.24. Haller B: Die Anatomie von Siphonaria gigas, Less., eines

opisthobranchen Gasteropoden. Arb Zool Inst Wien Zool Triest 1893,10:71-100.

25. Köhler A: Beiträge zur Anatomie dur Gattung Siphonaria. ZoologischeJahrbücher, Abteilung Anatomie 1894, 7:1-92.

26. Dayrat B, Tillier S: Goals and limits of phylogenetics. The Euthyneurangastropods. In Molecular Systematics and Phylogeography of Mollusks. Editedby: Lydeard C, Lindberg DR. Washington, DC: Smithsonian Books;2003:161-184.

27. Marcus E, Marcus E: On Siphonaria hispida. Universidade de São Paulo.Faculdade de Filosofia, Ciências e Letras. Boletim de Zoologia 1960,23:107-139.

28. Paul MC, Zubía E, Ortega MJ, Salvá J: New polypropionates fromSiphonaria pectinata. Tetrahedron 1997, 53(6):2303-2308.

29. Cimino G, Ghiselin MT: Chemical Defense and the Evolution of OpisthobranchGastropods San Francisco: California Academy of Sciences; 2009.

30. Boore JL, Brown WM: Big trees from little genomes: Mitochondrial geneorder as a phylogenetic tool. Curr Opin Genet Dev 1998, 8:668-674.

31. Sokolov EP: An improved method for DNA isolation frommucopolysaccharide-rich molluscan tissues. J Moll Stud 2000,66:573-575.

32. Boore JL, Macey JR, Medina M: Sequencing and comparing wholemitochondrial genomes of animals. In Molecular Evolution: Producingthe Biochemical Data. Part B. Methods in Enzymology. Volume 395. Editedby: Zimmer EA, Roalson E. Burlington, Massachusetts: Elsevier;2005:311-348.

33. Ewing B, Hillier L, Wendl MC, Green P: Base-calling of automatedsequencer traces using phred. I. Accuracy assessment. Genome Res 1998,8:175-185.

34. Gordon D: Viewing and editing assembled sequences using Consed. InCurrent Protocols in Bioinformatics. Edited by: Baxevanis AD, Davison DB.New York: John Wiley 2004:11.2.1-11.2.43.

35. Tamura K, Dudley J, Nei M, Kumar S: MEGA4: Molecular EvolutionaryGenetics Analysis (MEGA) software version 4.0. Mol Biol Evol 2007,24:1596-1599.

36. Dawid IB: Evolution of mitochondrial DNA sequence in Xenopus. Dev Biol1972, 29:139-151.

37. Brown WM, George M, Wilson AC: Rapid evolution of animalmitochondrial DNA. Proc Natl Acad Sci USA 1979, 76(4):1967-1971.

38. Attardi G: Animal mitochondrial DNA: an extreme example of geneticeconomy. Int Rev Cytol 1985, 93:93-145.

39. Song HJ, Sheffield NC, Cameron SL, Miller KB, Whiting MF: Whenphylogenetic assumptions are violated: base compositionalheterogeneity and among-site rate variation in beetle mitochondrialphylogenomics. Syst Entom 2010, 35(3):429-448.

40. Milne I, Wright F, Rowe G, Marshal DF, Husmeier D, McGuire G: TOPALi:Software for Automatic Identification of Recombinant Sequences withinDNA Multiple Alignments. Bioinformatics 2004, 20:1806-1807.

41. Stamatakis A, RAxML-VI-HPC: Maximum likelihood-based phylogeneticanalyses with thousands of taxa and mixed models. Bioinformatics 2006,22:2688-2690.

42. Rota-Stabelli O, Yang Z, Telford M: MtZoa: A general mitochondrial aminoacid substitutions model for animal evolutionary studies. Mol Syst Evol2009, 52:268-272.

43. Drummond AJ, Rambaut A: BEAST: Bayesian evolutionary analysis bysampling trees. BMC Evol Biol 2007, 7:214.

44. Ronquist F, van der Mark P, Huelsenbeck J: Bayesian phylogenetics usingMrBayes. Theory. In The Phylogenetic Handbook: a practical approach tophylogenetic Analysis and Hypothesis testing. Edited by: Lemey P, Salemi M,Vandamme A-M. Cambridge: Cambridge University Press; 2009:210-236.

45. Shimodaira H, Hasegawa M: CONSEL: for assessing the confidence ofphylogenetic tree selection. Bioinformatics 2001, 17:1246-1247.

46. Aktipis SH, Giribet G, Lindberg DR, Ponder WF: Gastropod phylogeny: anoverview and analysis. In Phylogeny and Evolution of the Mollusca. Editedby: Ponder WF, Lindberg DR. Berkeley: University of California Press;2008:199-236.

47. Kurabayashi A, Ueshima R: Complete sequence of the mitochondrial DNAof the primitive opisthobranch gastropod Pupa strigosa: Systematicimplication of the genome organization. Mol Biol Evol 2000, 17(2):266-277.

48. Grande C, Templado J, Cervera JL, Zardoya R: The complete mitochondrialgenome of the nudibranch Roboastra europaea (Mollusca: Gastropoda)supports the monophyly of opisthobranchs. Mol Biol Evol 2002,19:1672-1685.

49. Knudsen B, Kohn AB, Nahir B, McFadden CS, Moroz LL: Complete DNAsequence of the mitochondrial genome of the sea-slug, Aplysiacalifornica: conservation of the gene order in Euthyneura. Mol PhylogenetEvol 2006, 38:459-469.

50. Rumpho ME, Worful JM, Lee J, Kannan K, Tyler MS, Bhattacharya D,Moustafa A, Manhart JR: Horizontal gene transfer of the algal nucleargene psbO to the photosynthetic sea slug Elysia chlorotica. Proc NatlAcad Sci USA 2008, 105(46):17867-17871.

51. Terrett JA, Miles S, Thomas RH: Complete DNA Sequence of themitochondrial genome of Cepaea nemoralis (Gastropoda: Pulmonata). JMol Evol 1996, 42:160-168.

52. DeJong RJ, Emery AM, Adema CM: The mitochondrial genome ofBiomphalaria glabrata (Gastropoda: Basommatophora), intermediatehost of Schistosoma mansoni. J Parasitol 2004, 90:991-997.

53. Feldmeyer B, Hoffmeier K, Pfenninger M: The complete mitochondrialgenome of Radix balthica (Pulmonata, Basommatophora), obtained bylow coverage shotgun next generation sequencing. Mol Phylogenet Evol2010, 57:1329-1333.

54. Bandyopadhyay PK, Stevenson BJ, Cady MT, Olivera B, Wolstenholme DR:Complete mitochondrial DNA sequence of a Conoidean gastropod,Lophiotoma, Xenoturris, cerithiformis: Gene order and gastropodphylogeny. Toxicon 2006, 48:29-43.

55. Simison WB, Lindberg DR, Boore JL: Rolling circle amplification ofmetazoan mitochondrial genomes. Mol Phylogenet Evol 2006, 39:562-567.

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 14 of 15

56. Bandyopadhyay P K, Stevenson BJ, Ownby J-P, Cady MT, Watkins M,Olivera BM: The mitochondrial genome of Conus textile, coxI-coxIIintergenic sequences and Conoidean evolution. Mol Phylogenet Evol 2008,46(1):215-223.

57. Ki J-S, Lee Y-M, Jung S-O, Horiguchi T, Cho H-S, Lee J-S: Mitochondrialgenome of Thais clavigera (Mollusca: Gastropoda): Affirmation of theconserved, ancestral gene pattern within the mollusks. Mol PhylogenetEvol 2010, 54:1016-1020.

58. Maynard BT, Kerr LJ, McKiernan JM, Jansen ES, Hanna PJ: MitochondrialDNA sequence and gene organization in Australian backup abaloneHaliotis rubra (Leach). Mar Biotechnol 2005, 7:645-658.

59. Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R: DNA primers foramplification of mitochondrial cytochrome c oxidase subunit I fromdiverse metazoan invertebrates. Mol Mar Biol Biotechnol 1994, 3:294-299.

60. Palumbi S, Martin A, Romano S, McMillan WO, Stice L, Grabowski G: TheSimple Fool’s Guide to PCR Version 2.0 University of Hawaii: Department ofZoology and Kewalo Marine Laboratory; 1991.

61. Boore JL, Brown WM: Mitochondrial genomes of Galathealinum,Helobdella, and Platynereis: Sequence and gene arrangementcomparisons indicate that Pagonophora is not a phylum and Annelidaand Arthropoda are not sister taxa. Mol Biol Evol 2000, 17:87-106.

doi:10.1186/1471-2148-11-295Cite this article as: White et al.: Ten new complete mitochondrialgenomes of pulmonates (Mollusca: Gastropoda) and their impact onphylogenetic relationships. BMC Evolutionary Biology 2011 11:295.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit

White et al. BMC Evolutionary Biology 2011, 11:295http://www.biomedcentral.com/1471-2148/11/295

Page 15 of 15